# 7 day rolling mean of cases and deathscov19_ind['New_cases_smoothed']=cov19_ind['New_cases'].rolling(window=7).mean()cov19_ind['New_deaths_smoothed']=cov19_ind['New_deaths'].rolling(window=7).mean()
/tmp/ipykernel_81473/937387499.py:2: SettingWithCopyWarning:
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead
See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
cov19_ind['New_cases_smoothed'] = cov19_ind['New_cases'].rolling(window=7).mean()
/tmp/ipykernel_81473/937387499.py:3: SettingWithCopyWarning:
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead
See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
cov19_ind['New_deaths_smoothed'] = cov19_ind['New_deaths'].rolling(window=7).mean()
In [5]:
daily_cases=cov19_ind['New_cases'].to_numpy()daily_deaths=cov19_ind['New_deaths'].to_numpy()daily_cases_smoothed=cov19_ind['New_cases_smoothed'].to_numpy()daily_deaths_smoothed=cov19_ind['New_deaths_smoothed'].to_numpy()days=cov19_ind['Date_reported'].to_numpy()fig=plt.figure(figsize=(10,5))plt.plot(days,daily_cases,lw=1.8,ls='-',color='k',alpha=0.4,label='Daily cases')plt.plot(days,daily_cases_smoothed,lw=1.8,ls='-',color='k',alpha=0.9,label='Daily cases 7-day avg.')plt.axvspan(81,149,alpha=0.2,color='tab:blue',label='National Lockdown 2020')plt.plot(days,daily_deaths,lw=1.8,ls='-',color='tab:red',alpha=0.4,label='Daily deaths')plt.plot(days,daily_deaths_smoothed,lw=1.8,ls='-',color='tab:red',alpha=0.9,label='Daily deaths 7 day avg.')plt.axvspan(519,529,alpha=0.2,color='tab:green',label='National Lockdown 2021')plt.xticks(rotation=45);plt.gca().xaxis.set_major_locator(mdates.MonthLocator(interval=1))plt.xlabel('Time',fontsize=15)plt.ylabel('Frequency',fontsize=15)plt.title('COVID-19 Cases in India',fontsize=20,loc='left')plt.legend(frameon=False,ncol=2,loc='upper right',bbox_to_anchor=[1.05,1.20])plt.savefig('./images/fig1.png',dpi=300)
In [6]:
cumulative_cases=cov19_ind['Cumulative_cases'].to_numpy()cumulative_deaths=cov19_ind['Cumulative_deaths'].to_numpy()days=cov19_ind['Date_reported'].to_numpy()fig=plt.figure(figsize=(10,5))plt.plot(days,cumulative_cases,lw=1.8,ls='-',color='k',alpha=0.8,label='Cumulative cases')plt.axvspan(81,149,alpha=0.2,color='tab:blue',label='National Lockdown 2020')# plt.plot(days, cumulative_deaths, lw = 1.8, ls = '-', color = 'tab:red', alpha = 0.8, label = 'Cumulative deaths')plt.axvspan(519,529,alpha=0.2,color='tab:green',label='National Lockdown 2021')plt.xticks(rotation=45);plt.gca().xaxis.set_major_locator(mdates.MonthLocator(interval=1))plt.xlabel('Time',fontsize=15)plt.ylabel('Cumulative Frequency',fontsize=15)plt.title('COVID-19 cumulative cases in India',fontsize=20,loc='center')plt.legend(frameon=False,ncol=1,loc='lower right')plt.savefig('./images/fig2.png',dpi=300)
Analysing the variations in Aerosol Optical Depth¶
Here the changes in AOD are analysed and to visualise the affect of CoVID lockdown on Aerosol content in the atmosphere in the Indian subcontinent which is a good proxy for pollutant concentrations.
Souce of the data -> Level-1 and Atmosphere Archive & Distribution System Distributed Active Archive Center or LAADS DAAC.
The data is based on the retreivals from the VIIRS instrument onboard the SNPP | JPSS-1 satellite and is a radiometer.
In [7]:
# importing the necessary libraries for geospatial analysisimportxarrayasxrimportcartopy.crsasccrsimportcartopy.featureascfeatureimportnetCDF4importdask
I want to import the data in bulk but each file does not contain the time dimension, so, it is added manually to each file using the following function in the preprocess argument of xarray's open_mfdataset function. And then I replace the manual dates with the dates of the actual files
aerosol types for land and ocean combined data set. 0=dust(land+ocean), 1=smoke, 2=high altitude smoke, 3=pyrocumulonimbus clouds, 4=non-smoke fine mode, 5=mixed(land+ocean), 6=background(land+ocean maritime), 7=fine dominated
histograms of retrieved aerosol type over land and ocean. Indices (counting from 0) 0=dust(land+ocean), 1=smoke, 2=high altitude smoke, 3=pyrocumulonimbus clouds, 4=non-smoke fine mode, 5=mixed(land+ocean), 6=background(land+ocean maritime), 7=fine dominated
modal value of retrieved aerosol type over land and ocean. 0=dust(land+ocean), 1=smoke, 2=high altitude smoke, 3=pyrocumulonimbus clouds, 4=non-smoke fine mode, 5=mixed(land+ocean), 6=background(land+ocean maritime), 7=fine dominated
standard deviation of aerosol optical thickness at 490, 550 670, 865, 1240, 1610, and 2250 nm over ocean
Array
Chunk
Bytes
4.32 GiB
1.73 MiB
Shape
(2557, 7, 180, 360)
(1, 7, 180, 360)
Count
10228 Tasks
2557 Chunks
Type
float32
numpy.ndarray
description :
VIIRS Deep Blue Level 3 daily aerosol data, 1x1 degree grid
comment :
Data are the arithmetic mean of all VIIRS Deep Blue/SOAR L2 data located in each grid element after filtering by confidence flag. Source are the best_estimate SDS in L2 files.
The data is obtained for the time period of 2015-2021. the purpose of downloading the long term data was to do a anomaly analysis of the yeas 2019 (pre-covid) and 2020 (post-covid) against the long term 2015-2021 mean
aerosol types for land and ocean combined data set. 0=dust(land+ocean), 1=smoke, 2=high altitude smoke, 3=pyrocumulonimbus clouds, 4=non-smoke fine mode, 5=mixed(land+ocean), 6=background(land+ocean maritime), 7=fine dominated
histograms of retrieved aerosol type over land and ocean. Indices (counting from 0) 0=dust(land+ocean), 1=smoke, 2=high altitude smoke, 3=pyrocumulonimbus clouds, 4=non-smoke fine mode, 5=mixed(land+ocean), 6=background(land+ocean maritime), 7=fine dominated
modal value of retrieved aerosol type over land and ocean. 0=dust(land+ocean), 1=smoke, 2=high altitude smoke, 3=pyrocumulonimbus clouds, 4=non-smoke fine mode, 5=mixed(land+ocean), 6=background(land+ocean maritime), 7=fine dominated
standard deviation of aerosol optical thickness at 490, 550 670, 865, 1240, 1610, and 2250 nm over ocean
Array
Chunk
Bytes
4.32 GiB
1.73 MiB
Shape
(2557, 7, 180, 360)
(1, 7, 180, 360)
Count
10228 Tasks
2557 Chunks
Type
float32
numpy.ndarray
description :
VIIRS Deep Blue Level 3 daily aerosol data, 1x1 degree grid
comment :
Data are the arithmetic mean of all VIIRS Deep Blue/SOAR L2 data located in each grid element after filtering by confidence flag. Source are the best_estimate SDS in L2 files.
levels=np.arange(0,0.85,0.05)fig,ax=plt.subplots(1,1,figsize=(7,6),subplot_kw={'projection':ccrs.PlateCarree()})cf=ax.contourf(x,y,mean_pentad_aod,cmap='turbo',levels=levels,transform=ccrs.PlateCarree())plot_background(ax)norm=BM(levels,256)fig.colorbar(m.cm.ScalarMappable(norm=norm,cmap='turbo'),ax=ax, \
orientation='vertical',shrink=0.9,aspect=20,pad=0.05,label='AOD')plt.xlabel('Longitude',fontsize=13)plt.ylabel('Latitude',fontsize=13)plt.title('2015-2021 Mean AOD over India',fontsize=15)plt.savefig('./images/fig3.png',dpi=300)
Plotting the MAM 2019 and 2020 AOD mean values and difference in them
In [33]:
levels=np.arange(0,1.15,0.05)fig,axarr=plt.subplots(1,3,figsize=(21,5),subplot_kw={'projection':ccrs.PlateCarree()})axlist=axarr.flatten()cf0=axlist[0].contourf(x,y,aod_ind_2019_mam.mean(dim='time').to_numpy(),cmap='turbo',levels=levels,transform=ccrs.PlateCarree())# axlist[0].set_xlabel('Longitude', fontsize=13)# axlist[0].set_ylabel('Latitude', fontsize=13)axlist[0].set_title('2019 MAM season mean AOD',fontsize=15)cf1=axlist[1].contourf(x,y,aod_ind_2020_mam.mean(dim='time').to_numpy(),cmap='turbo',levels=levels,transform=ccrs.PlateCarree())# axlist[1].set_xlabel('Longitude', fontsize=13)# axlist[1].set_ylabel('Latitude', fontsize=13)axlist[1].set_title('2020 MAM season mean AOD',fontsize=15)norm=BM(levels,256)fig.colorbar(m.cm.ScalarMappable(norm=norm,cmap='turbo'),ax=axlist[:2], \
orientation='vertical',shrink=0.9,aspect=20,pad=0.025,label='AOD')levels=np.arange(-0.6,0.65,0.05)cf2=axlist[2].contourf(x,y,aod_ind_2020_mam.mean(dim='time').to_numpy()-aod_ind_2019_mam.mean(dim='time').to_numpy(),cmap='RdBu_r',levels=levels,transform=ccrs.PlateCarree())# axlist[2].set_xlabel('Longitude', fontsize=13)# axlist[2].set_ylabel('Latitude', fontsize=13)axlist[2].set_title('Change in AOD',fontsize=15)norm=BM(levels,256)fig.colorbar(m.cm.ScalarMappable(norm=norm,cmap='RdBu_r'),ax=axlist[2], \
orientation='vertical',shrink=0.9,aspect=20,pad=0.025,label='AOD change')# fig.set_constrained_layout_pads(w_pad=0.1, h_pad=0.1, hspace=0., wspace=0.)foraxinaxlist:plot_background(ax)# plt.tight_layout()plt.savefig('./images/fig4.png',dpi=300)
Plotting the anomaly of 2019 MAM and 2020 MAM season against the long term (2015-2021) mean.
In [17]:
levels=np.arange(-0.6,0.7,0.1)fig,axarr=plt.subplots(1,2,figsize=(14,6),subplot_kw={'projection':ccrs.PlateCarree()})axlist=axarr.flatten()cf0=axlist[0].contourf(x,y,aod_ind_2019_mam.mean(dim='time').to_numpy()-mean_pentad_aod,cmap='RdBu_r',levels=levels,transform=ccrs.PlateCarree())# axlist[0].set_xlabel('Longitude', fontsize=13)# axlist[0].set_ylabel('Latitude', fontsize=13)axlist[0].set_title('2019 MAM season AOD anomaly',fontsize=15)cf1=axlist[1].contourf(x,y,aod_ind_2020_mam.mean(dim='time').to_numpy()-mean_pentad_aod,cmap='RdBu_r',levels=levels,transform=ccrs.PlateCarree())# axlist[1].set_xlabel('Longitude', fontsize=13)# axlist[1].set_ylabel('Latitude', fontsize=13)axlist[1].set_title('2020 MAM season AOD anomaly',fontsize=15)norm=BM(levels,256)fig.colorbar(m.cm.ScalarMappable(norm=norm,cmap='RdBu_r'),ax=axlist[:2], \
orientation='vertical',shrink=0.9,aspect=20,pad=0.025,label='AOD anomaly')# fig.set_constrained_layout_pads(w_pad=0.1, h_pad=0.1, hspace=0., wspace=0.)foraxinaxlist:plot_background(ax)# plt.tight_layout()plt.savefig('./images/fig5.png',dpi=300)
Analysing the time series analysis for the year 2018 and 2019 (pre-covid) and 2020 (post-covid) and associate kernel density distrbutions
fig=plt.figure(figsize=(16,5))plt.subplot(121)plt.plot(np.arange(1,366,1),aod_ind_time_series_2018,lw=1.8,color='tab:red',alpha=0.8,label='AOD 2018')plt.plot(np.arange(1,366,1),aod_ind_time_series_2019,lw=1.8,color='tab:blue',alpha=0.8,label='AOD 2019')plt.plot(np.arange(1,367,1),aod_ind_time_series_2020,lw=1.8,ls='-',color='k',alpha=0.8,label='AOD 2020')plt.xlabel('Day of the year',fontsize=15)plt.ylabel('AOD',fontsize=15)plt.legend(frameon=False,ncol=1,fontsize=13)plt.title('Time series of AOD over the years',fontsize=20)plt.grid(axis='y',alpha=0.1)plt.subplot(122)sns.kdeplot(aod_ind_time_series_2018,lw=1.5,color='tab:red',label='AOD 2018',alpha=0.8)sns.kdeplot(aod_ind_time_series_2019,lw=1.5,color='tab:blue',label='AOD 2019',alpha=0.8)sns.kdeplot(aod_ind_time_series_2020,lw=1.5,ls='-',color='k',label='AOD 2020',alpha=0.8)plt.xlabel('AOD',fontsize=15)plt.ylabel('Density',fontsize=15)plt.legend(frameon=False,ncol=1,fontsize=13)plt.title('Kernel Density Estimate',fontsize=20)plt.savefig('./images/fig6.png',dpi=300)
It is evident from the analysis of AOD retrievals that the post-covid year of 2020 and the lockdown period has been associated with lower AQI when compared with the AOD values of pre-covid years, be it 2019 or 2020
To further confirm that this reduction in the AQI of the 3 sample years is not by change, a non-parametric Friedman's Chi square test is employed on the three annual AOD samples of 2018, 2019 and 2020 -
The p-value << 0.001 indicating rejecting the hypothesis that this change in AOD variability is by chance and is statistically significant. The Friedman' test is similar to ANOVA test but it is advantageous in the sense that it is non-parametric and doesn't assume any particular distribution of the samples beforehand.
A similar analysis if further carried out for individual pollutants of CO, HCHO and NO2¶
The Source of the data is from the TROPOMI satellite retrievals which have been preprocessed and available at TROPOMI QA4ECV. The CO and HCHO data were available in netcdf formats but the NO2 data was in ESRI grid format for download as monthly means.
2020-11-18 22:57:04 - Product generated with FORLI v20151001 Level 2 CO data.
summary :
This dataset contains Level 3 carbon monoxide gridded total columns from IASI observations. This dataset was generated by LATMOS under the auspices of AC SAF. Quicklooks and data access are provided by AERIS.
source :
ULB-LATMOS IASI Level 2 carbon monoxide (CO) data version 20151001
references :
Reference to the CO retrieval: FORLI radiative transfer and retrieval code for IASI, J. Quant. Spectrosc. Ra., 113, 1391���1408, https://doi.org/10.1016/j.jqsrt.2012.02.036, 2012.
levels=np.arange(0,16,1)fig,axarr=plt.subplots(1,3,figsize=(21,5),subplot_kw={'projection':ccrs.PlateCarree()})axlist=axarr.flatten()cf0=axlist[0].contourf(x,y,no2_2019_mam_ind.to_numpy(),cmap='turbo',levels=levels,transform=ccrs.PlateCarree(),extend='max')# axlist[0].set_xlabel('Longitude', fontsize=13)# axlist[0].set_ylabel('Latitude', fontsize=13)axlist[0].set_title('2019 MAM season mean',fontsize=15)cf1=axlist[1].contourf(x,y,no2_2020_mam_ind.to_numpy(),cmap='turbo',levels=levels,transform=ccrs.PlateCarree(),extend='max')# axlist[1].set_xlabel('Longitude', fontsize=13)# axlist[1].set_ylabel('Latitude', fontsize=13)axlist[1].set_title('2020 MAM season mean',fontsize=15)norm=BM(levels,256,extend='max')fig.colorbar(m.cm.ScalarMappable(norm=norm,cmap='turbo'),ax=axlist[:2], \
orientation='vertical',shrink=0.9,aspect=20,pad=0.025,label=r'Total Column NO2 ($\times 10^{15}$ molecules cm$^{-2}$)')levels=np.arange(-1,1.1,0.1)cf2=axlist[2].contourf(x,y,no2_2020_mam_ind.to_numpy()-no2_2019_mam_ind.to_numpy(),cmap='RdBu_r',levels=levels,transform=ccrs.PlateCarree(),extend='both')# axlist[2].set_xlabel('Longitude', fontsize=13)# axlist[2].set_ylabel('Latitude', fontsize=13)axlist[2].set_title('Change in Total Column NO2',fontsize=15)norm=BM(levels,256,extend='both')fig.colorbar(m.cm.ScalarMappable(norm=norm,cmap='RdBu_r'),ax=axlist[2], \
orientation='vertical',shrink=0.9,aspect=20,pad=0.025,label=r'Total Column NO2 change ($\times 10^{15}$ molecules cm$^{-2}$)')# fig.set_constrained_layout_pads(w_pad=0.1, h_pad=0.1, hspace=0., wspace=0.)foraxinaxlist:plot_background(ax)# plt.tight_layout()plt.savefig('./images/fig9.png',dpi=300)
It can be concluded here that the concentrations of tropospheric column CO have marginally increased while the changes in HCHO concentrations are not clearly significant. But for NO2 there is a significant reduction in the concentrations especially in the Southern Indian Peninsula and the Indo-Gangetic Plain.
Using Google Earth Engine, a similar analysis to compare the 2019 and 2020 concentrations of NO2, SO2 and Aerosol Index with the TROPOMI Sentinel-5P product is done.¶
Water quality was analysed using the Google Earth engine and Sentinel-2 MSI : MultiSpectral Instrument, Level-2A bands to calculate the value of Normalised Difference Chlorophyll Index (NDCI), which varies between -1 and 1 and higher the value, higher is the turbidity and poor is the quality of water¶
The analysis is done for Ganges in the Bhagalpur region for pre-lockdown (March 2019) and post-lockdown (April-May 2019) period.
Make this Notebook Trusted to load map: File -> Trust Notebook
The analysis of Ganges water quality in the Bhagalpur region through which Ganges passes shows a higher turbidity (poor water quality) in the pre-lockdown period and relatively better water quality in the post-lockdown period This showcases that the CoVID-19 lockdown did improve the water quality in the Bhagalpur district.
Further analysis over other branches of Ganga and other rivers can be carried out but are left out from this analysis largely due to time constraint.
For analysis of impact of CoVID lockdown on vegetation we used MERRA tropospheric reanalysis data for the satellite era produced by NASA Global Modeling and Assimilation Office (GMAO) using the Goddard Earth Observing System Model (GEOS).¶
The monthly mean data used to anlyse the differences in the 2019 and 2020 MAM seasons, which was the lockdown period in the year 2020. The variables of Surface soil wetness, available soil water storage and soil infiltration rate were anaylsed to determine the vegetation productivity by proxy. Higher values of each indicate better agricultural productivity and higher fertility for vegetation.
In [97]:
ds=xr.open_mfdataset('./data/MERRA/*.nc4')
In [98]:
ds
Out[98]:
<xarray.Dataset>
Dimensions: (time: 84, lat: 61, lon: 57)
Coordinates:
* lat (lat) float64 5.0 5.5 6.0 6.5 7.0 ... 33.0 33.5 34.0 34.5 35.0
* lon (lon) float64 65.0 65.62 66.25 66.88 ... 98.75 99.38 100.0
* time (time) datetime64[ns] 2015-01-01T00:30:00 ... 2021-12-01T00:...
Data variables: (12/24)
TSOIL5 (time, lat, lon) float32 dask.array<chunksize=(1, 61, 57), meta=np.ndarray>
TSOIL3 (time, lat, lon) float32 dask.array<chunksize=(1, 61, 57), meta=np.ndarray>
SPWATR (time, lat, lon) float32 dask.array<chunksize=(1, 61, 57), meta=np.ndarray>
RUNOFF (time, lat, lon) float32 dask.array<chunksize=(1, 61, 57), meta=np.ndarray>
TSOIL4 (time, lat, lon) float32 dask.array<chunksize=(1, 61, 57), meta=np.ndarray>
SWLAND (time, lat, lon) float32 dask.array<chunksize=(1, 61, 57), meta=np.ndarray>
... ...
GRN (time, lat, lon) float32 dask.array<chunksize=(1, 61, 57), meta=np.ndarray>
TSOIL1 (time, lat, lon) float32 dask.array<chunksize=(1, 61, 57), meta=np.ndarray>
TSOIL6 (time, lat, lon) float32 dask.array<chunksize=(1, 61, 57), meta=np.ndarray>
WCHANGE (time, lat, lon) float32 dask.array<chunksize=(1, 61, 57), meta=np.ndarray>
EVPTRNS (time, lat, lon) float32 dask.array<chunksize=(1, 61, 57), meta=np.ndarray>
TELAND (time, lat, lon) float32 dask.array<chunksize=(1, 61, 57), meta=np.ndarray>
Attributes: (12/32)
History: Original file generated: Mon Mar 23 22...
Filename: MERRA2_400.tavgM_2d_lnd_Nx.201501.nc4
Comment: GMAO filename: d5124_m2_jan10.tavg1_2d...
Conventions: CF-1
Institution: NASA Global Modeling and Assimilation ...
References: http://gmao.gsfc.nasa.gov
... ...
Contact: http://gmao.gsfc.nasa.gov
identifier_product_doi: 10.5067/8S35XF81C28F
RangeBeginningTime: 00:00:00.000000
RangeEndingTime: 23:59:59.000000
DODS_EXTRA.Unlimited_Dimension: time
history: 2022-08-29 03:55:02 GMT Hyrax-1.16.3 h...
levels=np.arange(0,1.1,0.1)fig,axarr=plt.subplots(1,3,figsize=(21,5),subplot_kw={'projection':ccrs.PlateCarree()})axlist=axarr.flatten()cf0=axlist[0].contourf(x,y,surface_soil_wetness_2019_mam.mean(dim='time').to_numpy(),cmap='YlGnBu',levels=levels,transform=ccrs.PlateCarree())# axlist[0].set_xlabel('Longitude', fontsize=13)# axlist[0].set_ylabel('Latitude', fontsize=13)axlist[0].set_title('2019 MAM season mean',fontsize=15)cf1=axlist[1].contourf(x,y,surface_soil_wetness_2020_mam.mean(dim='time').to_numpy(),cmap='YlGnBu',levels=levels,transform=ccrs.PlateCarree())# axlist[1].set_xlabel('Longitude', fontsize=13)# axlist[1].set_ylabel('Latitude', fontsize=13)axlist[1].set_title('2020 MAM season mean',fontsize=15)norm=BM(levels,256)fig.colorbar(m.cm.ScalarMappable(norm=norm,cmap='YlGnBu'),ax=axlist[:2], \
orientation='vertical',shrink=0.9,aspect=20,pad=0.025,label='Surface soil wetness')levels=np.arange(-0.5,0.6,0.1)cf2=axlist[2].contourf(x,y,surface_soil_wetness_2020_mam.mean(dim='time').to_numpy()-surface_soil_wetness_2019_mam.mean(dim='time').to_numpy(),cmap='BrBG',levels=levels,transform=ccrs.PlateCarree())# axlist[2].set_xlabel('Longitude', fontsize=13)# axlist[2].set_ylabel('Latitude', fontsize=13)axlist[2].set_title('2020 and 2019 MAM difference',fontsize=15)norm=BM(levels,256)fig.colorbar(m.cm.ScalarMappable(norm=norm,cmap='BrBG'),ax=axlist[2], \
orientation='vertical',shrink=0.9,aspect=20,pad=0.025,label='Surface soil wetness change')# fig.set_constrained_layout_pads(w_pad=0.1, h_pad=0.1, hspace=0., wspace=0.)foraxinaxlist:plot_background(ax)# plt.tight_layout()plt.savefig('./images/fig10.png',dpi=300)
In [127]:
levels=np.arange(0,1600,100)fig,axarr=plt.subplots(1,3,figsize=(21,5),subplot_kw={'projection':ccrs.PlateCarree()})axlist=axarr.flatten()cf0=axlist[0].contourf(x,y,available_water_storage_2019_mam.mean(dim='time').to_numpy(),cmap='YlGnBu',levels=levels,transform=ccrs.PlateCarree())# axlist[0].set_xlabel('Longitude', fontsize=13)# axlist[0].set_ylabel('Latitude', fontsize=13)axlist[0].set_title('2019 MAM season mean',fontsize=15)cf1=axlist[1].contourf(x,y,available_water_storage_2020_mam.mean(dim='time').to_numpy(),cmap='YlGnBu',levels=levels,transform=ccrs.PlateCarree())# axlist[1].set_xlabel('Longitude', fontsize=13)# axlist[1].set_ylabel('Latitude', fontsize=13)axlist[1].set_title('2020 MAM season mean',fontsize=15)norm=BM(levels,256)fig.colorbar(m.cm.ScalarMappable(norm=norm,cmap='YlGnBu'),ax=axlist[:2], \
orientation='vertical',shrink=0.9,aspect=20,pad=0.025,label='Available Water storage ($kg m^{-2}$)')levels=np.arange(-200,220,20)cf2=axlist[2].contourf(x,y,available_water_storage_2020_mam.mean(dim='time').to_numpy()-available_water_storage_2019_mam.mean(dim='time').to_numpy(),cmap='BrBG',levels=levels,transform=ccrs.PlateCarree(),extend='both')# axlist[2].set_xlabel('Longitude', fontsize=13)# axlist[2].set_ylabel('Latitude', fontsize=13)axlist[2].set_title('2020 and 2019 MAM difference',fontsize=15)norm=BM(levels,256,extend='both')fig.colorbar(m.cm.ScalarMappable(norm=norm,cmap='BrBG'),ax=axlist[2], \
orientation='vertical',shrink=0.9,aspect=20,pad=0.025,label='Available water storage change ($kg m^{-2}$)')# fig.set_constrained_layout_pads(w_pad=0.1, h_pad=0.1, hspace=0., wspace=0.)foraxinaxlist:plot_background(ax)# plt.tight_layout()plt.savefig('./images/fig11.png',dpi=300)
In [134]:
levels=np.arange(0,10.5,0.5)*10**(-5)fig,axarr=plt.subplots(1,3,figsize=(21,5),subplot_kw={'projection':ccrs.PlateCarree()})axlist=axarr.flatten()cf0=axlist[0].contourf(x,y,infil_rate_2019_mam.mean(dim='time').to_numpy(),cmap='YlGnBu',levels=levels,transform=ccrs.PlateCarree())# axlist[0].set_xlabel('Longitude', fontsize=13)# axlist[0].set_ylabel('Latitude', fontsize=13)axlist[0].set_title('2019 MAM season mean',fontsize=15)cf1=axlist[1].contourf(x,y,infil_rate_2020_mam.mean(dim='time').to_numpy(),cmap='YlGnBu',levels=levels,transform=ccrs.PlateCarree())# axlist[1].set_xlabel('Longitude', fontsize=13)# axlist[1].set_ylabel('Latitude', fontsize=13)axlist[1].set_title('2020 MAM season mean',fontsize=15)norm=BM(levels,256)fig.colorbar(m.cm.ScalarMappable(norm=norm,cmap='YlGnBu'),ax=axlist[:2], \
orientation='vertical',shrink=0.9,aspect=20,pad=0.025,label='Inflitration rate ($kg m^{-2} s^{-1}$)')levels=np.arange(-3,3.2,0.2)*10**(-5)cf2=axlist[2].contourf(x,y,infil_rate_2020_mam.mean(dim='time').to_numpy()-infil_rate_2019_mam.mean(dim='time').to_numpy(),cmap='BrBG',levels=levels,transform=ccrs.PlateCarree(),extend='both')# axlist[2].set_xlabel('Longitude', fontsize=13)# axlist[2].set_ylabel('Latitude', fontsize=13)axlist[2].set_title('2020 and 2019 MAM difference',fontsize=15)norm=BM(levels,256,extend='both')fig.colorbar(m.cm.ScalarMappable(norm=norm,cmap='BrBG'),ax=axlist[2], \
orientation='vertical',shrink=0.9,aspect=20,pad=0.025,label='Infiltration rate change ($kg m^{-2} s^{-1}$)')# fig.set_constrained_layout_pads(w_pad=0.1, h_pad=0.1, hspace=0., wspace=0.)foraxinaxlist:plot_background(ax)# plt.tight_layout()plt.savefig('./images/fig12.png',dpi=300)
The results of the analysis clearly indicate an increase in all three proxies in the lockdown season of 2020. Hence, an improvement in soil fertility owing to lockdown is evident
NDVI analysis was also carried out in the Sundarban region (dense forrest) and Haryana and Punjab (known for agricultural productivity) to determine the impact of CoVID on vegetaion in these regions using satellite imagery.¶
Data source - the satellite imagery was downloaded from the USGS Earth Explorer website, and we used Landsat-8 Bands 4 (red) and 5 (nir) for the analysis and calculation of NDVI
defget_bands(location,year):filepath='./data/USGS/'+location+f'/{year}'try:forfileinos.listdir(filepath):if'B4'infile:img=rasterio.open(f'{filepath}/{file}')red=img.read(1).astype('f4')elif'B5'infile:img=rasterio.open(f'{filepath}/{file}')nir=img.read(1).astype('f4')exceptFileNotFoundError:print('Directory does not exist, try a different directory.')print(filepath)returnred,nir
The results of our analysis compare the 2019 MAM and 2020 MAM seasons and highglight the the Sundarban densely forrested area NDVI increased during lockdown in 2020, while the agricultural lands in Haryana and Punjab actually saw a marginal devline in NDVI
It is clear that there has been a significant decline in AOD and aerosol concentrations in the Indian subcontinent during and post-lockdown. But, not all pollutant species show a decline such as CO and HCHO.
Water quality in the Bhagalur region imporoved post-lockdown owing largely to reduced industrial activties. Further analysis over other sectors of Ganga or other rivers can be carried out to further improve the results
The soil quality and fertility proxies indicate an improvement in soil water storage and infiltration during CoVID lockdown.
While the vegetation density seems to improve in the dense Sundarban region, the agricultural lands of Haryana and Punjab actually show a decline in NDVI and vegetation density.
